I’d like to take a very fast depth picture of a moving object, from two ZED2i devices. In this case, timestamp difference between two pictures (from two devices) should be less than 1ms. I try to run
grab for two devices parallelly in two threads. I compare
getTimestamp(sl::TIME_REFERENCE::IMAGE).getMicroseconds() from two cameras. The result is at least 2ms and usually is something around 10-15ms.
I use parameters for grab:
depth_mode = sl::DEPTH_MODE::PERFORMANCE; coordinate_units = sl::UNIT::MILLIMETER; depth_maximum_distance = 1400; camera_resolution = sl::RESOLUTION::HD720; camera_fps = 30;
Is it possible to make those
grabs closer to each other? E.g. there could be same way of capturing pictures from devices without processing. When we decide that timestamps are close enough, we would process them on GPU to prepare a depth picture and so on.